19results about How to "Improve defect detection performance" patented technology

The invention provides a PCBA surface defect detection method and system based on image recognition, and the method comprises the steps: collecting and constructing a multi-class PCBA overall defect image data set, and carrying out the preprocessing; constructing a pre-training neural network model, and classifying the preprocessed PCBA overall defect image data set; constructing a local target identification method based on a YOLOX network according to an overall defect classification result; and carrying out local defect detection on the PCBA surface based on a local target identification method. According to the method, the overall classification and local detection strategies are comprehensively applied to reduce the burden of the model, and the detection precision and speed are improved. And an attention mechanism is introduced to improve the performance and generalization ability of the model. Data enhancement, consistency processing, classification labeling and the like are carried out on an image data set, and the defect detection effect of the model is improved. And multiple optimization strategies such as structure improvement, loss function improvement and pruning lightweight are adopted for the network model, so that the image target detection performance is effectively improved.

The invention relates to a generator rotor surface defect detection method and system based on cross-modal fusion. The method comprises the following steps: acquiring an infrared image and a visible light image of the outer surface of a generator rotor magnetic pole; and respectively inputting the infrared image and the visible light image into a preset double-flow target detection model, and outputting a defect detection result of the outer surface of the generator rotor magnetic pole. The double-flow target detection model is composed of a backbone network, a PaFPN neck and a detection head; the backbone network comprises an infrared branch, a visible light branch and an image fusion module. According to the invention, two parallel branches of the backbone network of the double-flow target detection model are used to carry out feature extraction on the infrared image and the visible light image, and cross-modal feature fusion is carried out on the features of two modals through the image fusion module based on cross attention, so that mutual reference and supplement of information between different modals are realized. The fused image features maintain the advantages of respective modalities and improve the overall defect detection effect at the same time.

The invention provides a rolled steel surface defect detection method based on an improved group convolution and feature enhancement fusion network of a multi-head attention mechanism, and the method comprises the steps: obtaining a rolled steel surface defect image data set; training a defect detection model according to the steel rolling surface defect image data set based on a grouping space multi-scale convolution module and a feature enhancement fusion network; and inputting a to-be-detected image into the defect detection model to obtain a target rolled steel surface defect detection result. According to the steel rolling surface defect detection method based on the multi-head attention mechanism improved group convolution and feature enhancement fusion network, a group space multi-scale convolution module and a feature enhancement fusion network improved YOLOv7 target detection model are introduced, and a pre-trained defect detection model is obtained; the pre-training defect detection model is trained based on the training sample to obtain the defect detection model, the image needing to be detected is input into the defect detection model, the defect detection result is obtained, and the model defect detection performance is improved.

The embodiment of the invention provides a GAN-based ship beam surface defect detection data augmentation method, which comprises the following steps: acquiring a part of ship beam surface defect data set, and dividing the ship beam surface defect data set into a training set and a test set; a conditional generative adversarial network is adopted, an object generation network is constructed, and variables adopted by the conditional generative adversarial network are type labels of the ship beam surface defects; constructing an image conversion network, combining the defect image generated by the object generation network with the defect-free real image to generate a defective picture, and enabling the generated image to be similar to the target image; training an object generation network, generating a part of defect objects by using the constructed object generation network, and training an image conversion network by using the generated defect objects and real defect objects in the ship beam surface defect data set; a random noise and defect category label is input into a trained ship beam surface defect generation model, and a corresponding ship beam surface defect picture with the category label is generated.

The invention provides a surface defect detection method for aero-engine blades, and the method comprises the steps: sequentially shooting each blade, and splicing the blades into a panoramic image; due to the fact that the images are partially overlapped, the images can be sequentially spliced into the panoramic image according to the shooting sequence on the basis of the overlapped parts. The panoramic image not only contains enough defect information, but also can fuse a plurality of blades into the same image, so that the problem that the blades are difficult to match with the corresponding blades after the defects are identified is greatly reduced. Inputting the panoramic image into a recognition model added with a space-frequency dynamic convolution module, a Mangban module and a joint attention mechanism; and finally, outputting defect parts of all blades of the engine to finish identification. On the basis of the scheme, all defects on all the blades can be identified only by carrying out an identification process on the identification model once; the process has enough detection accuracy and does not need a large amount of calculation; and relatively high identification and blade quality detection efficiency can be kept.

The present invention discloses a battery coating defect detection device, which belongs to the field of defect detection. The battery coating defect detection device comprises a transmission component and a detection component, wherein the transmission component is used to transmit the battery pole piece to be detected; the detection component comprises an image sensor and a light source component, wherein the light source component comprises a light emitting component; the light emitting component is used to emit light inclined relative to the battery surface to the surface of the battery pole piece to generate a bright field; the image sensor is used to collect an image of the surface of the battery pole piece in the bright field. The battery coating defect detection device of the present invention can effectively highlight the defect characteristics, which is conducive to improving the defect detection effect.

An optical path device, an optical detection method and a detection device, the optical path device comprising a side incidence optical path assembly and a lens assembly, the side incidence optical path assembly comprising a detection light source and an expander mirror set, the detection light source providing an incident light beam, the expander mirror set being configured to spread and shape the incident light beam in a first direction to form parallel first direction detection light beams and to spread and shape the incident light beam in a second direction to form non-parallel second direction detection light beams; the first direction detection light beams exiting the lens assembly to form a light spot on a surface to be detected of an object to be detected; the second direction detection light beams non-parallelly entering a surface of the lens assembly and having a beam size, the second direction detection light beams exiting the lens assembly to the surface to be detected and forming the light spot. The present application provides a detection optical path avoiding a central aperture of the lens assembly, which can effectively reduce the risk of lens damage and achieve more reliable and efficient wafer defect detection.

The invention discloses a defect detection device for PP film production, which comprises a rack device, a detection device, a detection device and a control device, the rack device is provided with a carrier roller, and the rack device is provided with a driving belt for driving the carrier roller to rotate; the polygonal mirror is arranged on the outer side of the carrier roller in an arc shape; the machine frame device comprises two sets of first machine frames, two sets of synchronous assemblies are arranged on the outer sides of the first machine frames, the two sets of synchronous assemblies are symmetrically arranged relative to the carrier roller, and industrial cameras are arranged on the synchronous assemblies and used for shooting PP films reflected on the polygon mirror. The synchronous assembly further comprises a first gear, the first gear is fixedly connected with the end of the carrier roller, an intermittent assembly is arranged on the first rack, a base is fixedly connected to the first rack, a second gear is rotationally connected to the base, and a fixing assembly used for fixing an industrial camera is arranged on the second gear. And the reset assembly is arranged on the second gear.

The present invention discloses an eddy current detection circuit, method, system, storage medium and terminal, which belong to the field of eddy current non-destructive detection technology. The invention mainly utilizes a differential bridge and a transformer conditioning circuit to form a dual-path differential probe, and utilizes the signal parameters generated by the first differential signal and the second differential signal under their respective alternating magnetic fields to analyze the relationship between the amplitude, phase parameters and lift-off respectively generated under the alternating magnetic field, and performs multi-parameter signal fusion, which can suppress lift-off. The dual-path differential output has a high detection capability for defects when the lift-off changes greatly.

The invention provides a flaw detection method and device, computer equipment and a storage medium. The method comprises the following steps: mapping all second feature data of a workpiece training sample to nodes corresponding to a graph structure; according to a preset random walk rule, carrying out walk in the mapped graph structure to generate a node sequence; obtaining a first vector representation of a plurality of target workpiece samples in the node sequence; determining the same feature data between any two target workpiece samples in the plurality of target workpiece samples, and enabling the minimum mean square error of the same feature data to be minimum so as to adjust a plurality of second vector representations in each target workpiece sample; forming a new first vector representation of the target workpiece sample by using the plurality of adjusted second vector representations, and performing multi-round iteration to obtain a workpiece flaw detection model; and determining whether the to-be-detected workpiece sample has flaws or not through the workpiece flaw detection model. The workpiece flaw detection accuracy can be greatly improved.

According to the semi-supervised wood defect detection method based on the generative adversarial network provided by the invention, effective information in an unlabeled wood defect image which is easy to obtain is effectively explored by adopting an edge discriminator network algorithm and a segmentation network algorithm of a semi-supervised learning framework of the generative adversarial network GAN; the semi-supervised framework uses a DeepLabV3 + semantic segmentation network as a generator; using a full convolutional network as a discriminator, inputting an image segmentation mask pair and outputting a corresponding confidence map, and judging whether the segmentation mask is true or false; the edge discriminator network and the segmentation network perform adversarial training in a mutual game mode, when the discriminator reaches the required discrimination capability through stage training, the generator generates a segmentation result of a label-free image and an image label pair formed by corresponding original images, and then inputs the segmentation result and the image label pair into the discriminator; a label part which is verified to be true through a discriminator is used as a supervision signal, and a data set is added for training of a segmentation network; the reliability and segmentation performance of the detection model can be improved.

The present application discloses an image processing method, a defect detection method, and related devices and storage media, the method comprising: obtaining multiple original images of a target object; wherein the multiple original images are obtained by an image acquisition device acquiring the surface of the target object under different light source conditions, and the different light source conditions include at least one of the light source type and the light source irradiation direction; using the multiple original images, at least two channel images corresponding to different defect types are obtained; wherein at least one channel image is obtained by performing enhancement processing on at least one original image corresponding to the defect type; splicing at least two channel images to obtain a multi-channel image; wherein the multi-channel image is used for defect detection of the target object. Through the above method, the present application can amplify the defect features of various types of defects to provide better quality images for subsequent defect detection of the target object, thereby improving the defect detection effect of various types of defects.

This disclosure provides an AOI imaging system based on a TDI line scan camera for image acquisition of a display panel under inspection, relating to the field of display manufacturing. The system includes a TDI line scan camera, an illumination system, and a control system. The TDI line scan camera includes a first TDI line scan acquisition channel and a second TDI line scan acquisition channel. The illumination system includes a first illumination optical path and a second illumination optical path with different illumination effects. The control system is configured to: respond to an image acquisition trigger signal received by the AOI imaging system, control the first TDI line scan acquisition channel and the second TDI line scan acquisition channel to periodically alternate exposure, while simultaneously controlling the first illumination optical path and the second illumination optical path to periodically alternate illumination; and the first illumination optical path is illuminated simultaneously with the exposure of the first TDI line scan acquisition channel, and the second illumination optical path is illuminated simultaneously with the exposure of the second TDI line scan acquisition channel, so as to achieve the output of two illumination effects of the display panel under inspection in a single scan.

The utility model discloses an automatic identification device for panel glass surface defects, which comprises a detection platform, a light source system, an image acquisition module, an image processing unit and an adjustable mechanical support, the detection platform is provided with a three-dimensional adjusting seat, and the light source system comprises a multispectral light source assembly and a light path adjusting mechanism. The multispectral light source assembly integrates a visible light source, an infrared light source and an ultraviolet light source through an optical fiber array. Through fusion of multispectral light source integration, dynamic light path adjustment, precision mechanical positioning and edge computing technologies, an excitation-acquisition-processing full-link closed-loop system is constructed, and the bottlenecks of a traditional detection device in the aspects of multi-defect type adaptability, dynamic detection precision, environment anti-interference capability and the like are solved.

The invention discloses a chip defect detection system and method based on visual identification, and relates to the technical field of chip defect detection.The chip defect detection method comprises the steps that S10, the surface of an epitaxial wafer is partitioned, and the internal area of each edge closed line in a grain boundary edge image corresponds to one partition; s20, predicting the defect index of each partition; s30, the resistivity uniformity of each partition is predicted; and S40, carrying out screening treatment on the epitaxial wafer with defects. According to the method, the surface of the epitaxial wafer is subjected to partition treatment, it is guaranteed that the infrared light spectrum obtained through the infrared Fourier transform spectrometer can better reflect the defect condition of each partition, the production quality of the epitaxial wafer is evaluated through multiple physical characteristics of the epitaxial wafer in combination with the resistivity uniformity of each partition, and the production efficiency of the epitaxial wafer is improved. And the defect detection precision of the system on the epitaxial wafer is further improved.

The invention relates to the technical field of robots, in particular to an expandable multi-joint universal intelligent robot which comprises a robot body arranged at the top of a first conveyor, the robot body comprises a controller, a rotating mechanism, a screening mechanism and an expanding mechanism, and the rotating mechanism comprises a rotating shaft and a driving assembly. The screening mechanism comprises an industrial camera, a telescopic assembly, an angle adjusting assembly and a grabbing assembly, the expanding mechanism comprises a collecting box, a distance sensor, a traction assembly and a pushing assembly, and the industrial camera, the distance sensor, the driving assembly, the telescopic assembly, the angle adjusting assembly, the grabbing assembly and the traction assembly are all electrically connected with the controller. The expandable multi-joint universal intelligent robot not only has a detection function, but also has a screening function, and workpieces with defects and workpieces without defects after detection can be quickly separated and treated separately without causing confusion.

The invention relates to the technical field of textile fabrics, in particular to a knitted fabric broken yarn defect detection system and method based on artificial intelligence. The method comprises the following steps: firstly, collecting knitted fabric images containing broken yarn defects, and constructing a target domain data set; secondly, acquiring defect images and labels of different knitted fabrics, and constructing a source field data set; performing pre-training on the OfficientNet by using the source field data set to obtain source field features; and inputting the target domain data set into the pre-trained OfficientNet to obtain a target domain feature. Secondly, transmitting the source domain feature and the target domain feature to a domain classifier, and outputting a domain label vector; thirdly, a defect detection loss function is designed and comprises a plurality of loss items, and all the loss items are added through weighting; and finally, the OfficientNet and the domain classifier are subjected to alternate updating. And outputting the broken yarn defect type of the knitted fabric image through the softmax layer according to the output of the OfficientNet. According to the invention, the problem of insufficient data volume of the defect image of the knitted fabric can be solved, and the defect detection accuracy is further improved.

The utility model relates to a braid detection device, in particular to a braid tension adjusting system in intelligent braid detection equipment, which is characterized by further comprising a tension feedback adjusting device, the tension feedback adjusting device comprises a tension sensing device and a feedback control unit, the tension sensing device is mounted at a braid taking position between a driving shaft M1 and a driving shaft M2, and the feedback control unit is connected with the tension feedback adjusting device. Wherein the rotating speed of one of the driving shaft M1 and the driving shaft M2 is a set fixed value, the other driving shaft is a variable speed driving shaft, the signal output end of the tension sensing device is connected to the feedback control unit, and the signal control end of the feedback control unit is connected with the power source end of the variable speed driving shaft. The device has the advantages that the device is suitable for different types of braids and the same type of braids with different widths and elasticity, the operation is convenient, and the efficiency is greatly improved; the operation stability is improved; the defect detection rate in visual detection is improved, and the deviation between a defect position indication point and an actual defect occurrence position is effectively reduced; and tension adjustment can be achieved in an automatic control mode, and operation is easy and convenient.

The present invention belongs to the technical field of welding tooling, and discloses a photoelectric sensor welding tooling with defect detection function and its working method, including a base, a top seat and a clamping seat, the bottom surface of the top seat is fixedly connected with a mounting sleeve, and the inner surface of the mounting sleeve is fixedly sleeved with a distribution component. The present invention drives the composite detection part to move from bottom to top along the curved trajectory through the arc rod in the curved telescopic part when gas is introduced, and gradually moves from the bottom of the side of the workpiece to the top, so as to perform a comprehensive detection on the side and top of the workpiece, and cooperate with the composite detection part in a synchronous rotating state, and the gap exhaust moves down and resets during the rotation process, so that the arc trajectory in the rotating state is formed on the outside of the workpiece. The rise and fall of the arc trajectory is realized on the outside of the workpiece after welding, and the comprehensive detection of the outer side and top surface of the workpiece after welding is realized, which greatly improves the scope of defect detection after welding, and the actual detection range is wide, the speed is fast, the effect is good, and the final quality of defect detection after welding is improved.